Marijuana is the most widely used illegal drug in the United States. Although marijuana use had decreased substantially since the late 1970s, it has risen dramatically since 1992 among teenagers and young adults. One possible factor contributing to this upsurge is the widespread belief that there are no chronic adverse health effects of regular marijuana use. In November 1996, California voters passed Proposition 215, which would legalize the use of marijuana for medicinal purposes. Despite scientific debate about the benefits of marijuana treatment, the public's perception seems to be that the potential benefits outweigh the risks. Nevertheless, several lines of evidence from biochemical, cellular, tissue and animal studies provide a biologically plausible basis for the hypothesis that marijuana is a risk factor for respiratory-tract cancers. Thus far, however, there is no epidemiologic evidence for this association, primarily because of the long induction/latency of human carcinomas and the infrequent use of marijuana in the general U.S. population before the late '60s. The major objective of the proposed 5-year project is to estimate the effects of marijuana use on the risk of lung cancer and upper-aerodigestive-tract (UAT) cancers among residents of Los Angeles County, ages 18-57. Secondary objectives are to assess the interaction effects of marijuana and tobacco use, to estimate the effects of other factors for which the epidemiologic evidence is inconsistent or sparse, and to initiate a molecular study by obtaining tumor specimens for cases and buccal cells from cases and controls. The proposed design is a population-based case-control study involving 600 lung-cancer cases, 600 UAT-cancer cases (395 oral cavity and pharynx, 85 esophagus, and 120 larynx), and 1,200 controls. Histologically confirmed cases will be identified by the rapid ascertainment system of the USC Cancer Surveillance Program, a population-based SEER registry for Los Angeles County. Controls will be selected according to a prespecified algorithm from the neighborhoods of cases; one control will be matched to each case on age, sex, race/ethnicity, and neighborhood. The major source of data will be personal interviews conducted with all subjects in their homes. Information will be collected on a variety of factors: detailed history of marijuana, tobacco and alcohol use; occupational and environmental exposures (e.g., passive smoking); clinical factors (e.g., depression and family history of cancer); diet and other behaviors (e.g., fruits, vegetables, fats, and supplements); and sociodemographic factors (e.g., SES).